The aim of the project is to make the following contributions to our understanding of the phonology-syntax interface. First, the project investigates the syntax and prosody of complex syntactic structures – embedded clauses, dislocations, and questions. This has allowed us to test whether prosodic models developed for simple structures extend to more complex data. Careful phonetic analysis of the data has provided a more detailed description of systematic correlates of intonation and phrasal prosody, allowing each prosodic level to be better motivated. Analyzing the data in current syntactic models has allowed for a more fine-grained look at the syntactic factors that condition prosody and determine whether non-syntactic factors influence phrasal prosody. Moreover, this research aims to make an important contribution to our understanding of Bantu syntax in terms of description, typology and syntactic theory. Investigating the same structures in several Bantu languages, spoken over a geographically wide area, allows for a comparative study of the syntax and prosody of these structures that should be of interest both to Bantuists and to other linguistic researchers working on similar issues. To achieve these goals, we adopted the following methodology. Each year of the project we focused on one complex construction: embedded clauses, questions and dislocations. We developed questionnaires to elicit comparable data for these constructions for each of the languages included in the project. Fieldwork to Africa often had to be undertaken in order to provide sufficient data for analysis for some understudied languages. Phonetic analysis of some the data collected was undertaken at the Paris phonetics laboratory, while phonological and syntactic analyses were undertaken by the individual researchers to explore the implications of the data for current linguistic theories. At our annual workshops, the entire group met to compare results in each language, identify gaps in the data and prepare the next year’s questionnaire. Towards the end of the project, a new methodology was developed for analyzing the Embosi data, using language technology tools developed for French. Our results make a contribution to each of our targeted research areas. For the study of intonation, we show that intonational tune can be superimposed on lexical tone, leading to a change of tone register. For prosodic phrasing theory, we show that the attested levels of prosodic phrasing often do not match the predictions of the prosodic hierarchy, while the attested prosodic phrasing domains do not match the domains predicted by syntactic phase theory. An important goal of the project was to make new data and analyses of understudied constructions in understudied languages available to other researchers interested in issues in the phonology-syntax interface.